The present invention relates to an information processing apparatus and method, a program, and a recording medium. More particularly, the present invention relates to an information processing apparatus and method, a program, and a recording medium which properly reflect the change in user's preference and properly infer user's interest in newly added attributes, thereby recommending contents which is closer to user's interest.
As methods for recommending television programs (or contents), the initial interest registering, the viewing log utilizing, and the coordination filtering are known, for example.
In any of these schemes, an electronic program guide (EPG) and program information (or program metadata) on the Web are used for the original data. These three schemes depend on how to acquire user's preferential data to be matched with these pieces of information.
In the initial interest registering, at the starting of use, the user registers his/her favorite categories (drama, variety, etc.), favorite genre name (drama, music, etc.), and favorite television personalities, for example. Subsequently, matching is made with program metadata with the registered information used as keywords, thereby getting program names to be recommended.
In the viewing log utilizing, every time the user views a program, the program metadata associated with the viewed program is accumulated and, when the viewing log (or program metadata) is accumulated to a predetermined amount, this viewing log is analyzed to get program names to be recommended. Alternatively, instead of the viewing log, an operation log such as the preset recording or the start of recording done by the user may be used on devices in which recording is made on a hard disk drive, for example. In this case, the information, which highly reflects user's interest, can be obtained rather than vague information about programs.
In the coordination filtering, a matching is made between one user's viewing (or operation) log and another user's viewing log to acquire the viewing log of another user similar to the user concerned in viewing log. Then, program names which have not been viewed by the user concerned are obtained from among the programs viewed by another user similar (in favorite) to the user concerned in viewing log, thereby recommending the obtained program names not viewed by the user concerned.
By using the above-mentioned program recommendation schemes, it becomes possible to allow the recommendation of the programs in which the user is interested.
However, these program recommendation schemes present problems that the interest of user is extracted from program (contents) metadata (namely, a biased interest in television programs is obtained) and, because generally popular names are used owning to the structure of program metadata, similar contents tend to be recommended.
To overcome these problems, the applicant hereof proposed a technology for letting users evaluate how much they are interested in particular contents recommended for them and searching for particular contents on the basis of the results of the evaluation, thereby making recommendations (refer to Patent Document 1, for example).
[Patent document 1]
Japanese Patent Laid-open No. 2002-63206
However, the related-art technologies presuppose the recommendations on the basis of the initially registered data and the user information such as contents viewing log, so that it is difficult to reflect the change in user preference and infer the user's interest in newly added genre, for example.